Official Claude Code plugin marketplace for ScrapeUnblocker - fetch any web page from behind anti-bot protection (Cloudflare, DataDome, PerimeterX, Akamai, Shape) without writing a scraper.
In Claude Code:
/plugin marketplace add ScrapeUnblocker/claude-code-plugin
/plugin install scrapeunblocker@scrapeunblocker
You will be asked for your ScrapeUnblocker API key when the plugin is enabled. It is stored in your OS keychain, not in a settings file. Get a key at app.scrapeunblocker.com.
MCP tools - available to Claude automatically:
| Tool | What it does |
|---|---|
fetch_html |
Fully rendered HTML of any page, anti-bot bypassed |
fetch_parsed |
AI-parsed structured JSON (products, articles, listings) instead of raw markup |
google_search |
Organic Google results as structured data |
fetch_markdown |
A page as clean Markdown, usually 2-25% the size of the HTML - the cheap way to read a page |
search_and_read |
A Google search plus its top 1-5 result pages as Markdown, in one call |
list_elements |
A page's links, inputs and buttons with ready-to-use selectors |
browse_page |
Click, type, select and scroll on a page in a real browser, then return the HTML (not read-only: it acts on the website) |
list_plugins / run_plugin |
100+ ready-made site scrapers (Amazon, Zillow, Google Maps, YouTube, App Store...) that return clean JSON |
amazon_search, amazon_product, amazon_reviews, walmart_search, ebay_search, google_shopping, google_local, maps_place_details, google_images, youtube_search, youtube_transcript, zillow_search, indeed_jobs, trustpilot_reviews, booking_hotels |
The most-used site plugins as tools of their own |
Command:
/scrapeunblocker:scrape-url https://example.com
/scrapeunblocker:scrape-url https://example.com parsed
/scrapeunblocker:scrape-url https://example.com markdown
/scrapeunblocker:scrape-url https://example.com country=US
Skills - reference material Claude pulls in when it is writing scraping code for your project:
best-practices- endpoints, parameters, retries, choosing the right callpage-source- fetching rendered HTML or Markdown (markdown=true), render waits, country targetingparsed-data- structured extraction and how to fix a bad parse
Anything a normal fetch or curl returns as 403, a captcha, or an empty JS shell comes back as the real page. Requests go through a real browser and a proxy pool, so Claude can read pages it otherwise cannot: pricing pages, marketplaces, listings, SERPs.
- Docs: docs.scrapeunblocker.com
- MCP server (standalone, for other MCP clients):
scrapeunblocker-mcp - SDKs: Python, Node.js, Ruby, PHP
MIT